Seth Klarman, founder of Baupost Group, built it from $27 million into $22 billion - and his secret weapon is doing nothing for years at a time
his book "Margin of Safety" is out of print and sells used for over $1,000
he almost never talks in public
he just did - in one of the rarest interviews of his career
75-min on how the Oracle of Boston actually thinks about risk
bookmark & watch today
this is f*cking dangerous
build hedge fund using "loop engineering" that prints alpha 24/7 (Full Guide)
if I had this a year ago, I would've built my hedge fund in a week instead of a year
bookmark before someone takes it down
2026 might be the year where hedge fund level research becomes accessible to literally anyone.
Some random guy with Claude, a few regulatory PDFs and $12 of API credits can now do the kind of reverse engineering that used to require an actual team of analysts.
Meanwhile most people still think AI is for anime pictures.
This Jane Street / SEBI breakdown is one of the first articles in a while that actually feels like a glimpse of where things are going
More like, AI turning highly asymmetric institutional workflows into open source infrastructure.
Great article to read.
-> https://t.co/ccvL9xpJix
A multi-agent LLM framework that simulates a hedge fund team analyzing stocks and generating trading signals, with support for Anthropic, OpenAI, and local models.
-> https://t.co/XdVkHXZfYW
A Python library for parsing and analyzing all major SEC EDGAR filing types (13F, ADV, 10-K, 8-K, 13D) with a built-in Claude MCP server for natural language queries.
JP Morgan headhunted trader for ~$500K/year from Lehman Brothers - after he predicted the bank's collapse under $600B in debt
- he explain why Goldman Sachs hired him at 20 and gave him ~$100M in his first week
27-min video to understand what strategies he uses in his trading that made every tier-1 fund in the world want to hire him
$AMD CEO Lisa Su said at CES 2026 that the world may need 100x more compute over the next five years as AI workloads scale across industries.
This earnings cycle is proving the point:
• $AMZN AWS reaccelerated to 28% (fastest growth in 15 quarters) while Bedrock customer spend grew 170% QoQ & processed more tokens in Q1 than all prior years combined
• $MSFT AI business is now at a $37B run-rate growing 123% YoY
• $GOOGL Cloud grew 63% (now nearly half the size of AWS) & has a $243B backlog while GenAI products grew 800% YoY
• $MU & $SNDK margins show memory demand is still compounding WAY faster than supply with margins now higher than many software companies
The 100x compute framework looks less crazy after this earnings cycle.
AI Semiconductor Endgame 2026 (Part 1)
New Token Economics Computing Paradigm Shifts from GPU Compute to HBM
This article starts from the essence of GPU architectural evolution to address a question the market has long worried about:
Why must each GPU's HBM memory demand grow exponentially, and why won't this exponential growth in HBM demand stall?
It then derives the first principle of token economics under the current architecture: token throughput = HBM size × HBM BW (bandwidth)
It also discusses why the GPU ceiling is determined by HBM's two dimensions of progress.
The topic of HBM cyclicality has long been controversial. Optimists argue that AI-driven demand is much greater than before, but the market mainstream still believes that previous up-cycles also saw 20%+ annual demand growth — so what's different this time? AI doesn't change the fact that HBM, like traditional DRAM, has commodity attributes. Once capacity expansion at the demand peak meets a downturn, history will repeat itself. We can take the perspective of compute-chip architecture, start from first principles, and unpack and reason through this question:
why this time is genuinely different.
———————————————————————————————
History: The Era of CPU Compute
For a very long time, we lived in the era of CPU-dominated compute. The CPU's top-level KPI was performance — running faster — and so each generation of CPUs deployed every method imaginable to push benchmark scores higher. First it was rising clock frequencies, then it was architectural evolution: superscalar designs, and so on.
During this period, why didn't DDR need to advance technologically at high speed? DDR3 to DDR5 took a full 15 years.
Because in this era, DDR's role was purely auxiliary — and only weakly so. By industry experience, even doubling DDR speed would generally only raise CPU performance by less than 20%.
Why did improvements in DDR bandwidth and speed matter so little? Two reasons:
1. CPUs designed all kinds of architectural tricks to hide DDR latency — superscalar designs, wider issue widths, massive ROBs and register renaming to extract parallelism and hide latency, L1 caches, L2 caches — all of which weakened the demand for DDR bandwidth and speed.
2. CPU workloads don't have particularly demanding bandwidth requirements. For most everyday workloads — say, opening a webpage — DDR bandwidth is severely overprovisioned. Even cloud workloads often look the same.
In other words, in the CPU era, DDR bandwidth and speed didn't really matter. There was virtually no difference between DDR4 and DDR5 except in a handful of games — and even the JEDEC standard advanced slowly.
On top of that, only a small portion of any given app needs to permanently sit in DDR. Whatever is needed can be paged in from the hard drive on demand. App size grew slowly, and so DDR capacity demand grew slowly as well.
That's why, over the past decade, the average PC went from 7–8GB of DDR to about 23GB — only 3× growth in ten years.
This slow upgrade pace directly affected revenue. Capacity-based pricing was the main way of making money; speed improvements were just a technological upgrade that raised the unit price of capacity. With both of these dimensions advancing slowly, growth could only come from increases in PC/phone unit volumes.
So along both dimensions — bandwidth/speed and capacity — DRAM was always a “nice-to-have” appendage to the chip industry. The marginal utility of DDR upgrades was very low, and almost completely disconnected from the CPU era's top-level KPI.
———————————————————————————————
The Paradigm Shift: GenAI's Top-Level KPI
When we entered the era of GenAI large models, the computing paradigm shifted, and the top-level KPI changed fundamentally.
By the time GPUs evolved into AI inference engines, the top-level KPI was no longer compute alone (TOPS/FLOPS), as it had been for CPUs — it became the cost of a token. Specifically: overall token throughput per unit cost / per unit power.
A close second is token throughput speed — because in the agent era, many tasks have become serial, and token output speed has become a critical bottleneck for user experience.
This is exactly why Jensen invented the concept of the AI factory: to produce the most tokens at the lowest cost, while pushing token throughput speed as high as possible.
In the AI training era, Jensen's economics were TCO (Total Cost of Ownership): the more GPUs you buy, the more you save.
In the inference era, Jensen's token economics flip the logic:
AI inference has very healthy gross margins, so the logic now becomes: the NVIDIA GPU is the GPU that produces the cheapest token in the world, so the more you buy, the more you earn.
The top-level KPI has become a Pareto frontier: along the two dimensions of token throughput and token speed, optimize as far as possible.
Each generation of NVIDIA's token factory is essentially pushing the entire Pareto frontier up and to the right. This is the most important KPI of the AI inference era.
———————————————————————————————
From Token Throughput to HBM: The Core Logic Chain
Below is the most important logical chain of this article: how to start from the exponential growth of token throughput and derive that the ceiling bottleneck lies in the exponential growth of HBM size and HBM speed.
In the era of single-GPU inference with single-thread batch size = 1, token throughput had only one dimension: HBM bandwidth speed. Higher bandwidth = higher token throughput.
But once we entered the NVL72 era, inference is no longer single-GPU. It is a system-level token factory composed of 72 GPUs + 36 CPUs, designed to fully saturate HBM bandwidth and compute simultaneously, in pursuit of the ultimate token throughput.
Token throughput growth depends on two things: the number of requests batched simultaneously × the average token speed per request.
That is: batch size × token speed.
Take Rubin NVL72 as an example. At an average token speed of 100 tokens/s, processing 1,920 simultaneous requests yields a token throughput of 192,000 tokens/s. A Rubin NVL72 draws roughly 120kW (0.12MW), so per MW it can handle 1.6M tokens/s.
So we need to find ways to push both parameters up: batch size and average token speed. Their product is our top-level KPI — token throughput.
Parameter 1: Batch growth — bottleneck is HBM size
Every request in the batch carries its own KV cache, which has to live in HBM, with sizes ranging from a few GB to tens of GB. Because hot KV cache must be read at high frequency and high speed at any moment, it must reside in HBM. For a model with, say, 80 layers, every token generation step requires reading the KV cache 80 times from HBM.
As batch size grows, hot KV cache grows linearly.
And because the hot KV cache for every request in the batch must sit in HBM, HBM size must grow linearly with batch size.
Like an airport shuttle bus: the gate wants to move passengers to the plane as fast as possible. If HBM size is small, the shuttle is small, so you have to make extra trips.
Conclusion: batch size growth bottlenecks on HBM size growth.
Parameter 2: Average token speed per request — bottleneck is HBM bandwidth
The decode-phase speed of a large model bottlenecks on HBM bandwidth, because every token generated requires reading the activated weights and KV cache many times over.
The emergence of LPUs has, in cases where batch size isn't very large, moved the activated weights portion onto SRAM — but every generated token still requires many reads of the KV cache from HBM. The higher the HBM bandwidth, the faster each token is generated, in essentially linear correspondence.
Like the airport shuttle bus: HBM bandwidth is like the width of the door — wider doors mean passengers board faster.
The rest of the GPU's configuration is essentially adapted to support batch growth and to keep token compute speed in step with HBM growth. In some cases the GPU even spends excess compute to recover effective bandwidth (e.g., bandwidth compression techniques).
—-------
To return to the shuttle bus analogy:
• Shuttle bus cabin size = HBM Size (capacity): determines how many passengers can fit at once (i.e., how many requests' KV caches can sit in HBM simultaneously). Bigger cabin = more passengers (higher batch size) per trip. If the bus is too small, moving 100 people takes two trips — and total throughput suffers.
• Shuttle bus door width = HBM Bandwidth: determines how fast passengers get on and off. A wide door, and everyone piles on at once (decode/token generation is fast). A narrow door, and even with a giant cabin, people queue up and most of the time is spent boarding.
• Passenger throughput = cabin size × door-width-determined boarding speed.
—-------
At this point, we've logically derived the first principle of token-economics hardware demand:
Token throughput = HBM size × HBM Bandwidth
The top-level KPI of the AI inference era is highly dependent on progress along both HBM dimensions.
If we want to maintain 2× token throughput growth per generation, that means each generation of single GPU must grow HBM size × HBM BW speed by 2×!
This is the first time in history that HBM memory size can influence the top-level KPI — token throughput.
To validate this thesis, we can put NVIDIA's token throughput from A100 to Rubin Ultra on the same chart as HBM size × HBM BW speed.
What you find is that the two curves track each other startlingly closely on log axes.
HBM size × speed actually grows even faster than token throughput — which makes sense, because HBM defines the ceiling, and in practice utilization of that ceiling is very hard to push to 100%. Even if HBM size × HBM speed grew by 1,000×, with the supporting compute and architecture, it would be very hard to wring out the full 1,000× of headroom.
This curve isn't a coincidence — it's the necessary solution of system optimization.
throughput = batch × speed. This is the unavoidable first principle of token factory economics.
—-------
What about software? Won't software optimization reduce bandwidth demand? Reduce HBM demand?
This is an independent dimension from hardware. It's like asking: if software on a CPU runs faster after optimization, does that mean the CPU doesn't need to advance for ten years? After all, software is faster now.
If that were the case, would CPU vendors still make money? For a CPU vendor to survive, there's only one path: in standardized benchmarks, ignoring software optimization, every new CPU generation must score higher — otherwise it doesn't sell.
GPUs are exactly the same. How well software is optimized, and the requirement that the GPU's own token-throughput KPI must improve dramatically every year, are two separate things.
As long as token demand keeps growing, the pursuit of higher token throughput will not stop — and so neither will the pursuit of higher HBM size × HBM speed.
If HBM size and HBM speed were to slow down, Jensen would personally fly to the Big Three and pressure them to accelerate, because that ishis GPU ceiling. If the ceiling stops rising, can his GPU still sell?
Of course, NVIDIA also needs to wrack its brains to extract performance beyond the HBM ceiling through heterogeneous architectural angles. The LPU is a great example — it improved the Pareto frontier substantially from a different angle (the right-hand high-token-speed portion).
—--------------------
HBM memory has now bid farewell to that old era of drifting with the tide. On this one-way road paved by exponential demand, it has, in something close to a destined fashion, walked onto the central stage of the industry's epic.
When the inference paradigm's first principles evolve to this point, as long as Jensen still wants to sell GPUs, HBM must double — and it must double every generation. This is endogenous pressure from the supply side. It has nothing to do with AI demand, nothing to do with macro cycles, and nothing to do with the moods of the hyperscalers.
The only remaining question is this:
When demand has been physically locked into exponential growth, will the three players on the supply side — like they have for the past thirty years — once again drag themselves back into the mire of the cycle by their own hands?
You’re bored because you’re not doing side quests, man.
Life is more than just working and then throwing yourself into bed doing nothing.
Here are 50 side quests to complete:
Five Fundamental Truths:
1. Anything can happen.
2. You don’t need to know what is going to happen next in order to make money.
3. There is a random distribution between wins and losses for any given set of variables that define an edge.
4. An edge is nothing more than an indication of a higher probability of one thing happening over another.
5. Every moment in the market is unique.
Mark Douglas
This master A.I. prompt will help you make as much money as possible in 2026.
It gives the latest AI models instructions to become your financial assistant, analyze your portfolio, and then optimize for maximum return this year.
(bookmark this post to reference later)
_____
You are an elite global macro investor and portfolio manager whose sole objective is to maximize absolute returns in calendar year 2026, while explicitly managing downside risk and drawdowns.
I have uploaded my full investment portfolio, including account statements, holdings, allocations, transaction history, and relevant personal constraints (tax status, liquidity needs, restrictions).
Primary Objective
Optimize my portfolio to make as much money as possible in 2026, not to minimize volatility or track benchmarks. Risk is acceptable if it is intentional, asymmetric, and well-rewarded.
____
👉 These prompts are best used on Silvia, a free product we built that gives AI the context necessary to provide you with the best answers. Try it free: https://t.co/bMI7hLeciU
----
Step 1: Portfolio X-Ray
Analyze my current portfolio and summarize:
- Asset allocation by category (equities, fixed income, crypto, commodities, private assets, cash, derivatives, etc.)
- Geographic exposure
- Factor exposure (growth, value, momentum, size, duration, leverage, liquidity)
- Correlations between major holdings
- Concentration risk (top positions, themes, single-factor exposure)
- Tax characteristics (taxable vs tax-advantaged, short vs long-term gains)
- Liquidity profile (what can and cannot be rebalanced quickly)
- Identify hidden risks, redundancies, and unintentional bets.
Step 2: Macro & Regime Assumptions for 2026
Form explicit macro assumptions for 2026, including:
- Monetary policy direction (rates, liquidity, QE/QT)
- Inflation vs deflation risks
- Fiscal policy and deficits
- Economic growth or recession probabilities
- Market regime (risk-on, risk-off, volatility)
- Geopolitical or structural tail risks
- Clearly state base case, bull case, and bear case scenarios and assign probabilities.
Step 3: Identify the Highest-Conviction Opportunities
Across all investable assets, identify where the largest risk-adjusted and absolute return opportunities are likely to exist in 2026.
Include:
- Public equities (regions, sectors, styles)
- Individual stock themes or baskets
- Crypto assets and protocols (if applicable)
- Commodities and real assets
- Rates, currencies, or volatility trades
- Levered or derivative strategies (where appropriate)
- Rank opportunities by expected upside, timeframe, and conviction.
Step 4: Rebalancing Plan
Design a clear rebalancing plan that answers:
- What to reduce or exit (low expected return, poor asymmetry)
- What to add or increase
- Target weights for major positions
- Where leverage (explicit or implicit) makes sense
- How to structure entries (lump sum vs staged)
- Explain why each change improves 2026 expected returns.
Step 5: Capital Deployment & Timing
Provide a capital deployment plan addressing:
- Timing considerations for rebalancing
- Event-driven or catalyst-based positioning
- How to keep dry powder available
- When to take profits
- How to respond if thesis breaks
Step 6: Risk Management That Doesn’t Kill Returns
Define guardrails that protect against catastrophic loss without neutering upside:
- Maximum drawdown tolerance
- Position sizing rules
- Correlation limits
- Optional hedges (if justified)
- “Kill switches” for broken theses
Avoid generic diversification advice—be precise.
Step 7: Tax-Aware Optimization
Optimize for after-tax 2026 returns by considering:
- Tax-loss harvesting
- Asset location
- Holding period optimization
- Income vs capital gains tradeoffs
- Rebalancing friction
Explicitly note when ignoring taxes may be rational.
Step 8: Output Requirements
Deliver the final answer as:
- Executive Summary (1–2 pages equivalent)
- Target 2026 Portfolio Allocation Table
- Top 5 Return Drivers for 2026
- Primary Risks to the Plan
Checklist of Actions to Execute
Use clear language, tables, and bullet points. State assumptions clearly. If information is missing, list the exact data you need.
Guidelines
- Optimize for making money, not optics
- Assume the reader understands markets
- No generic financial-planning boilerplate
- Be decisive and explicit
- Legal, ethical, and compliant strategies only
The output should read like a confidential 2026 portfolio memo prepared by a hedge-fund CIO for a high-conviction investor.
____
👉 These prompts are best used on Silvia, a free product we built that gives AI the context necessary to provide you with the best answers. Try it free: https://t.co/bMI7hLeciU
----